Fine metabolic regulation in ruminants via nutrient–gene interactions:saturated long-chain fatty acids increase expression of genes involved inlipid metabolism and immune response partly through PPAR-a activation
Massimo Bionaz†, Betsy J. Thering† and Juan J. Loor*
Mammalian NutriPhysioGenomics, Division of Nutritional Sciences, 498 Animal Sciences Laboratory, Department of
Animal Sciences, University of Illinois, Urbana, IL 61801, USA
(Received 15 December 2010 – Revised 30 March 2011 – Accepted 12 April 2011 – First published online 6 July 2011)
Abstract
Madin–Darby Bovine Kidney cells cultured with 150mM of Wy-14 643 (WY, PPARa agonist) or twelve long-chain fatty acids (LCFA; 16 : 0,
18 : 0, cis-9–18 : 1, trans-10–18 : 1, trans-11–18 : 1, 18 : 2n-6, 18 : 3n-3, cis-9, trans-11–18 : 2, trans-10, cis-12–18 : 2, 20 : 0, 20 : 5n-3 and
22 : 6n-3) were used to uncover PPAR-a target genes and determine the effects of LCFA on expression of thirty genes with key functions
in lipid metabolism and inflammation. Among fifteen known PPAR-a targets in non-ruminants, ten had greater expression with WY,
suggesting that they are bovine PPAR-a targets. The expression of SPP1 and LPIN3 was increased by WY, with no evidence of a similar
effect in the published literature, suggesting that both represent bovine-specific PPAR-a targets. We observed the strongest effect on the
expression of PPAR-a targets with 16 : 0, 18 : 0 and 20 : 5n-3.When considering the overall effect on expression of the thirty selected
genes 20 : 5n-3, 16 : 0 and 18 : 0 had the greatest effect followed by 20 : 0 and c9t11–18 : 2. Gene network analysis indicated an overall
increase in lipid metabolism by WY and all LCFA with a central role of PPAR-a but also additional putative transcription factors. A greater
increase in the expression of inflammatory genes was observed with 16 : 0 and 18 : 0. Among LCFA, 20 : 5n-3, 16 : 0 and 18 : 0 were the most
potent PPAR-a agonists. They also affected the expression of non-PPAR-a targets, eliciting an overall increase in the expression of genes
related to lipid metabolism, signalling and inflammatory response. Data appear to highlight a teleological evolutionary adaptation of PPAR
in ruminants to cope with the greater availability of saturated rather than unsaturated LCFA.
Key words: Transcriptomics: Ruminants: Liver: PPAR response element
PPAR-a, PPAR-g and PPAR-b/d, which are a sub-class of the
nuclear hormone receptor superfamily, represent potential
molecular targets to prevent metabolic disorders associated
with lipid metabolism, glucose metabolism (particularly
PPAR-g) and immune function (i.e. have anti-inflammatory
potential)(1). In non-ruminants, and particularly in rodents,
PPAR-a is highly expressed in the liver and has a pivotal
role in increasing the oxidation of long-chain fatty acids
(LCFA)(2,3). A beneficial role of PPAR-a in reducing fatty
liver and overall liver inflammatory response has been
demonstrated in mice(4–6).
From a practical standpoint, the prevention of fatty liver in
peripartal dairy cows, with the associated metabolic issues
(e.g. ketosis), is without doubt a priority in order to prevent
impairment in performance(7). Fatty liver in peripartal cows
is caused by the surge of NEFA as a consequence of a
marked negative energy balance(7). Based on the data on
rodents, it has been proposed(7) that activation of PPAR-a
could improve the overall capacity of the peripartal bovine
liver to oxidise LCFA, thus preventing fatty liver. Except
for few preliminary data(8–12) there is a lack of information
available on the regulation of PPAR activity in ruminant cells
or tissues.
The activity of PPAR in non-ruminants is modulated
by LCFA. The potency of inducing transcription through
PPAR activation varies among the types of LCFA(1) and it
is dose-dependent(13). Other than preliminary evidence of
an effect of LCFA on ruminant PPAR activation(10,11), the
† These authors contributed equally.
*Corresponding author: Dr J. J. Loor, fax þ1 217 333 8286, email [email protected]
Abbreviations: ACOX1, acyl-coenzyme A oxidase 1; CD36, CD36 molecule; CLA, conjugated linoleic acid; CPT1A, carnitine palmitoyltransferase 1A; CTR,
control; FABP3, fatty acid binding protein 3; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reductase; HP, haptoglobin; LCFA, long-chain fatty acid;
LPIN1, lipin 1; MDBK, Madin–Darby Bovine Kidney cell; PPRE, PPAR response element; qPCR, quantitative real-time RT-PCR; SCD, stearoyl CoA
desaturase; SREBF1, sterol regulatory element binding transcription factor 1; WY, Wy-14 643.
British Journal of Nutrition (2012), 107, 179–191 doi:10.1017/S0007114511002777q The Authors 2011
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
transcriptomics effects associated with specific LCFA, both
dietary and rumen-derived saturated and unsaturated,
remain relatively unknown.
The specific objectives of the present study were to (1) find
reliable bovine PPAR-a targets among key metabolic genes
(mostly established PPAR-a target genes in non-ruminants)
after treatment with Wy-14 643 (WY; a potent specific
PPAR-a agonist in bovine endothelial cells(14,15)); (2) test the
effect of several LCFA on PPAR-a activation by measuring
the expression of the bovine-specific PPAR-a genes and (3)
investigate LCFA-specific effects, besides PPAR-a activation,
on networks among the genes measured.
Materials and methods
Fatty acid preparation, cell culture and treatments
Treatments included WY (270-198-M010, Alexis Biochemicals,
Lausen, Switzerland), palmitic acid (16 : 0; N-16-A, Nu-Chek
Prep, Inc., Elysian, MN, USA), stearic acid (18 : 0; N-18-A,
Nu-Chek Prep, Inc.), oleic acid (cis9–18 : 1; 1022, Matreya,
Pleasant Gap, PA, USA), trans10–18 : 1 (provided by Dr R. A.
Erdman, University of Maryland, College Park, MD, USA),
vaccenic acid (trans11–18 : 1; U-49-A, Nu-Chek Prep, Inc.),
linoleic acid (18 : 2; 215040050, Acros Organics, Morris
Plains, NJ, USA), rumenic acid (c9,t11CLA; no. 1245, Matreya),
trans10, cis12–18 : 2 (t10, c12CLA; no. 1249, Matreya), a-linole-
nic acid (18 : 3; no. 302820010, Acros Organics), phytanic acid
(20 : 0; no. 1195, Matreya), EPA (20 : 5n-3; no. N-20-A, NuChek
Prep, Inc.) and DHA (22 : 6n-3, no. 90 310, Cayman Chemical
Company, Ann Arbor, MI, USA). Fatty acids were saponified
using an equimolar concentration of NaOH and dissolved in a
final solution of 95 % ethanol to obtain a stock concentration
of 30 mM. WY was dissolved in 95 % ethanol.
Preliminary data demonstrated that Madin–Darby Bovine
kidney cells (MDBK) are a suitable in vitro model to test
PPAR activation(10,11). The MDBK cells were obtained from
ATCC (CCL-22, Manassas, VA, USA) at passage 110. A previous
investigation was conducted to characterise them and to set
optimal conditions for the present study(11). The results from
the previous investigation indicated 6 h of incubation, use
of LCFA not bound to albumin and addition of insulin to the
culture were the best conditions to measure maximal
expression of most genes of interest. Based on another pre-
liminary study(10), the present study was performed using
150mM each of LCFA and WY in order to directly compare
the potency in the activation of gene expression among
treatments. All treatments were administered in HyQw Mini-
mum Essential Media/Earle’s Balanced Salts (MEM/EBSS;
no. SH30024·02, HyClone, Logan, UT, USA) without fetal
bovine serum and containing bovine insulin (5 mg/l; no.
1882, Sigma, St Louis, MO, USA). In addition, an ethanol
control (CTR) to account for the ethanol effect (5 ml/l) and
only media were run. All treatments were run in triplicate.
After 6 h incubation the cells were harvested in 1 ml TRIzolw
reagent (Invitrogen, Carlsbad, CA, USA) and immediately
stored at 2808C until RNA extraction.
Transcripts measured
The description, main function(s) and sub-cellular location of the
products of the genes of interest are reported in Table S1
(additional file 1, available online at http://www.journals.
cambridge.org/bjn). Those genes were chosen partly because of
their potential as targets of PPAR-a in non-ruminant species, as
highlighted by Fig. S1 (additional file 1, available online at http://
www.journals.cambridge.org/bjn) and based on key functions in
LCFA uptake and trafficking (e.g. CD36 molecule (CD36), fatty
acid binding protein 3 (FABP3)), LCFA oxidation (e.g. acyl-coen-
zyme A oxidase 1 (ACOX1), carnitine palmitoyltransferase 1A
(CPT1A)), TAG synthesis (e.g. stearoyl-CoA desaturase (SCD),
lipin 1 (LPIN1)), cholesterol synthesis (e.g. 3-hydroxy-3-methyl-
glutaryl-coenzyme A reductase (HMGCR)), gene transcription
(e.g. sterol regulatory element binding transcription factor 1
(SREBF1)), gluconeogenesis (pyruvate carboxylase (PC)) and
immune response (including acute-phase reaction) (e.g. IL6, hap-
toglobin (HP)). We also characterised the response to LCFA of iso-
forms of novel genes, which have been demonstrated to be
crucial in bovine mammary TAG synthesis(16) and are targets(17)
or co-activators(18,19) of non-ruminant PPARa (acyl-CoA synthe-
tase long-chain family member and lipin isoforms, respectively).
Additional materials and methods
The details of RNA extraction and quantitative real-time RT-PCR
(qPCR), relative mRNA abundance between measured transcripts,
network development using Ingenuity Pathway Analysisw
(Ingenuity Systems Inc., Redwood City, CA, USA) and hierarchical
clustering among genes and treatments are reported in additional
file 1 (supplementary material available online http://www.
journals.cambridge.org/bjn). qPCR performance and primer
features are reported in Tables S2 and S3 (additional file 1, avail-
able online at http://www.journals.cambridge.org/bjn).
Minimum Information for publication of Quantitativereal-time PCR Experiments (MIQE) compliance ofquantitative real-time RT-PCR data
The qPCR data with all available information have been
submitted as an Real-time PCR Data Markup Language
(RDML)(20) file and, except for RNA integrity, all the other infor-
mation required by the MIQE guideline(21) are provided in the
main paper or in additional file 1 (supplementary material avail-
able online at http://www.journals.cambridge.org/bjn).
Statistical analysis
qPCR-normalised data are presented as log2-transformed fold-
change relative to CTR. In earlier statistical analysis all data
were transformed as fold-change relative to CTR and log2
transformed to normalise the data, to minimise the effects of
the outliers and to prevent a bias towards the treatments
with extremely large effects. The presence of possible outliers
was assessed by PROC REG of SAS (SAS Institute Inc., Cary,
NC, USA, release 9.0) on log2-transformed data and data
points with studentised residuals $2·5 were considered
M. Bionaz et al.180
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
outliers and excluded from the analysis. This final data set was
analysed using a generalised linear model (GLM). Treatment
was considered as a fixed effect and replicate as a random
effect. The multiple comparisons were corrected using
Tukey’s test. Significance was declared at P-corrected # 0·05
for all comparisons. Pearson’s correlation analysis was run
using the PROC CORR procedure of SAS.
Peroxisome proliferator-activated receptor response elementand three-dimensional structure modelling analyses
In order to uncover the potential PPAR response element (PPRE)
in the sequence (promoter þ coding sequence) of themeasured
genes we used the software RESearch(22). To evaluate the simi-
larity of three-dimensional structure of PPAR-a between species
and between PPAR isotypes we used the Swiss-Pdb Viewer(23).
Both methodologies are described in detail in additional file 1
(supplementary material available online at http://www.
journals.cambridge.org/bjn).
Results and discussion
Bovine PPAR-a target genes
Among several methods available to assess the activity of
PPAR-a, e.g. through firefly luciferase reporter, the use of
well-established agonists is a reliable and easy, albeit indirect,
alternative. In all mammalian cells tested to date(24), including
bovine(14,15), WY has been demonstrated to be a potent
PPAR-a activator and a weak PPAR-g activator(25). We used
WY as positive CTR in the present study.
Figs. S1 and S2 (additional file 1, available online at http://
www.journals.cambridge.org/bjn) report the percentage of
abundance among the measured genes and the known net-
works among all the genes measured. The networks generated
using both the Ingenuity Pathway Analysis Knowledge Base
and previous reports(26–28) encompass fifteen out of thirty
genes measured in the present study, whose expression has
been demonstrated to be specifically under control of PPAR-a
(i.e. these are downstream PPAR-a target genes). Among the fif-
teen known PPAR-a target genes in the non-ruminants that we
tested, nine genes including ACSL1, ACSL3, ANGPTL4, CD36,
CPT1A, FABP4, LPIN1, SCD and SREBF1 were significantly
up-regulated by treatment with WY (Figs. 1 and 2 and summary
in Fig. 3) confirming they also are PPAR-a targets in bovines
(bold arrows in Fig. 3). Among the remaining PPAR-a target
genes reported in non-ruminants, FABP3, DBI, ACOX1 and
UCP2 were not significantly affected by WY and HMGCS1 had
only a numerical increase in expression (Fig. 3). In addition,
in non-ruminants, PPAR-a controls the expression of its own
gene (i.e. PPARA)(26,29). Our data clearly indicated that 6 h incu-
bation with WY did not affect PPARA expression (Fig. 2). These
data support previous results from a 24 h time-course exper-
iment using MDBK cells(11). Expression of SPP1 and LPIN3
appeared to be induced by WY (Table 1 and Fig. 2) but they
have not been previously reported to be PPAR-a targets; thus,
they can be considered putative bovine-specific PPAR-a targets
(dashed lines in Fig. 3).
Among the fifteen known PPAR-a target genes in non-
ruminants (Fig. 3), ten (approximately 67 %) were confirmed
to be PPAR-a target genes in MDBK cells. A comparison
with other studies where hepatocytes from mice, humans or
rats were treated with WY(24,27) (see additional file 1 for
detailed discussion, supplementary material for this article
can be found at http://www.journals.cambridge.org/bjn)
revealed some overlap in response but a greater sensitivity
of MDBK to WY compared with non-ruminant hepatocytes.
Ligand-induced activation of PPAR-a in non-ruminants,
especially in rodents, leads to altered expression of many
genes associated with fatty acid oxidation as well as other
liver-specific functions(3,24,26). In this regard, besides the lack
in response of ACOX1, our data in MDBK suggest a potential
increase in the capacity for LCFA oxidation through greater
LCFA entry and transport into cytoplasm ( * CD36, * FABP4
and * ACSL1) as well as mitochondria ( * CPT1A; Figs. 1
and 3), all data that appear to support the findings in non-
ruminants. In addition, our data suggest that some of the
steps leading to the synthesis of TAG and cholesterol are
under the control of PPAR-a (Fig. 3). The control of the
expression of the genes related to desaturation, as well of
LPIN1 (30), by PPAR-a in mouse liver has been reported pre-
viously(26,31), but there are no reports, to our knowledge, of
an increased amount of intra- or extra-cellular quantity of
TAG via the activation of PPAR-a. We recognise that the
increase in the expression of one or few genes in a pathway
would probably not increase the total metabolic flux, as dis-
cussed previously(16,32). However, the induction of lipogenesis
as a consequence of PPAR-a activation seems more supported
by the increase in SREBF1 expression, which is a well-
established regulator of lipid synthesis in non-ruminants
(particularly de novo LCFA synthesis)(33). The increase in
expression of this gene after treatment with WY also has
been observed in mouse and human hepatocytes(24,28). The
biological meaning of this is not apparent, but reinforces pre-
vious findings from our laboratory using bovine mammary
cells, where it was evident that activation of PPARg after 6 h
of treatment with rosiglitazone increased SREBF1 expression
approximately twofold(34). Overall, it appears that bovine
SREBF1 is under control of PPAR isotypes. We cannot exclude
that the increase in expression of SREBF1 in the present study
was due to a partial activation of PPARg by WY(25). However,
recent data in HepG2 cells appear to support the idea that
PPARa through cross-talk with SREBP signalling(28) controls
lipid and cholesterol synthesis.
To evaluate a potential relationship between PPRE and
genes responsive to WY, we conducted an in silico analysis
that provided number, location and strength of PPRE in the
sequence of all measured genes(22). The analysis of PPRE
(see Tables S6 and S7, Figs. S3 and S4 and relative results
and discussion in additional file 1, supplementary material
for this article can be found at http://www.journals.
cambridge.org/bjn) provided some support for the gene
expression data as reflected by the fact that we found a greater
percentage of medium-strong PPRE in genes affected by WY
compared with the non-WY-sensitive genes (additional files 2
and 3, supplementary material for this article can be found
PPAR transcriptomics in bovine cells 181
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
at http://www.journals.cambridge.org/bjn). However, the
in silico approach for the prediction of PPAR target genes in
bovines appears to be weak. The system used to evaluate
the PPRE in the selected genes was developed for non-
ruminants(22) and a difference in PPRE response between
non-ruminant species has been demonstrated for ACOX1 (35,36).
An interesting outcome from the PPRE analysis was the
finding of a PPRE in bovine SPP1. This gene has not
been reported to be a PPAR-a target in non-ruminants but,
rather, its expression is down-regulated by the activation
of both PPAR-a(37) and PPAR-g(38). However, those results
were not confirmed by transcriptomics analysis in mice,
humans(24) or rats(27), where SPP1 expression was unchanged
after WY treatment. PPRE analysis in bovines clearly showed
that this gene presents, uniquely among all measured genes,
only two medium-strength PPRE for PPAR-a located far
up-stream relative to the transcription start site (probably con-
sidered distal) but none, except a weak PPRE, for the other
(B)
–2
0
2
4
6
8(A)
–1·0–0·50·00·51·01·52·0
(C)
–0·8–0·6–0·4–0·20·00·20·40·6
(D)
–1012345
f
e e,f
d,e d,ec,dc,d
b,cba
a
f
c,d,ed,ec,d,e c,d,e
c,d,e
b,c,d,e
b,c,db,c,da,b a,b,c
a
d,e ec,d,ed,ec,d,e b,c,d,e c,d,e b,c,d,e c,d,e
a,b,c,db,c,d
a,b,cab
a
c
b,cb,ca,b
a
b,c,d,e
b,cb,c,d
(F)
–1
0
1
2
3
4(E)
Log
2 fo
ld m
RN
A r
elat
ive
to C
TR
–0·50·00·51·01·52·02·53·0
(G)
–2–1012345 (H)
–2·0–1·5–1·0–0·50·00·51·0
g
fc,d,e
e,fd,e,f
a,bb,c,d a,b,c
a
d,e
bc,d
d,ec,d
ecb,c
a
eb,c
a
g
d
f,g
e
c
f,g
e,f
b
f,g
a
b
c,d,e b,c,da,b
e e e
g
a
(J)
–3–2–10123
(I)
–1
0
1
2
3
(K)
Treatments
M MCTR
WY
16:0
18:0
c9–18
:1
t10–
18:1
t11–1
8:118
:2
c9t11
–18:2
t10c
12–1
8:218
:320
:020
:522
:6CTR
WY
16:0
18:0
c9–18
:1
t10–
18:1
t11–1
8:118
:2
c9t11
–18:2
t10c
12–1
8:218
:320
:020
:522
:6
–1
0
1
2
3(L)
–1
0
1
2
3
cd,e,f
d,e,f
d d,ec c
b
d,e,fb,c,d b,c
e,f,g
d,e,f
e,f,g
b,c,d,ea,ba
h
a,b
b,c,ddc,d
a,b,c
b,c,d
a
c,d
a,b,c,d a,b,c,d c
bc c
b
d,e
a
bc
d,e,f f
a
f,g f,g g,h g,h
a
c,de e
d,e,f
e,f
c,d
a,b
Fig. 1. Effect of 150mM of Wy-14 643 (WY) or several long-chain fatty acid (LCFA) treatments for 6 h on the expression of selected genes related to lipid metab-
olism. Mean values with unlike letters were significantly different (P,0·05; Tukey’s corrected). (A) ACSL1, acyl-coenzyme A synthetase long-chain family member
1; (B) CD36, fatty acid translocase CD36 molecule (thrombospondin receptor); (C) FABP3, fatty acid-binding protein 3; (D) FABP4, fatty acid-binding protein 4;
(E) CPT1A, carnitine palmitoyltransferase 1A; (F) SCD, stearoyl-coenzyme A desaturase; (G) LPIN1, lipin 1; (H) DGAT1, diacylglycerol-O-acyltransferase homo-
log 1; (I) HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reductase; (J) HMGCS1, 3-hydroxy-3-methylglutaryl-coenzyme A synthase 1 (soluble); (K) SREBF2,
sterol regulatory element-binding transcription factor 2; (L) SREBF1, sterol regulatory element-binding transcription factor 1. M, media; CTR, control (ethanol).
M. Bionaz et al.182
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
two PPAR isotypes (Fig. S3 in additional file 1, supplementary
material for this article can be found at http://www.journals.
cambridge.org/bjn). Together, with the gene expression
results (Fig. 2), these data strongly support the notion that
bovine SPP1 is a specific target gene of PPAR-a in bovines.
We conducted an alignment and three-dimensional structure
analyses of the PPAR-a proteins with the purpose of evaluating
the potential differences in PPAR-a response to WY observed
between the species (see above). The conservation of amino
acid sequence of PPAR-a proteins between mice, humans and
bovines is .90 %, with 100 % conservation in the DNA-binding
domain, which interacts with the PPRE (Table S8 in the
additional file 1, supplementary material for this article can be
found at http://www.journals.cambridge.org/bjn). The remark-
able degree of homology at the protein level is indicative of a
high degree of functional conservation, which in turn suggests
that bovine PPAR-a should be able to bind to non-ruminant
PPRE. This has been clearly demonstrated by the successful acti-
vation of a luciferase construct with rat acyl-CoA oxidase PPRE
in bovine cells(39). The sequence homology of the ligand-bind-
ing domain, which constitutes the pocket for the entry and bind-
ing of agonists, is .90 % conserved between the three species,
with almost 98 % conservation between bovines and humans
(Table S8·2 in additional file 1, supplementary material for this
article can be found at http://www.journals.cambridge.org/
bjn), but the differences observed with the three-dimensional
protein structure of the ligand-binding domain (Figs. S5–S9
and related results and discussion in additional file 1, sup-
plementary material for this article can be found at http://
www.journals.cambridge.org/bjn) appear to shed some light
on the varying response of mice and bovines to PPAR-a agonists
despite the high degree of conservation of the primary structure.
The data indicated a larger and more neutral ligand pocket in
bovine compared with mouse PPAR-a. Detailed molecular ana-
lyses will have to be performed to determine the specific conse-
quences of the observed differences but might explain the
contrast in WY response between species and the poor agree-
ment between the PPRE analysis and our gene expression data.
(H)
–101234
d,e,f,gd,e,f,gg
b,c,d,e
b,c,d
a
f,g
b,c,db,c,d,e
bc,d,e,f b,c
e,f,g
(A)
–0·4–0·20·00·20·40·60·8
(B)
–1·5–1·0–0·50·00·51·01·52·0
(C)
–1012345
(D)
–1·2–1·0–0·8–0·6–0·4–0·20·00·20·4
b,cc,d,e
d,ed,e
c,d,e
e
a
b,c,da,b
a
a,b,c,d,e
a,b
e,f
a,b,c,d
a,b,c,d,ed,e,f
g
c,d,e,f
f
b
c,d,e
e
a,b,cb,c,d,e
b,c,d,e
(E)
Log
2 fo
ld m
RN
A r
elat
ive
to C
TR
–2
–1
0
1
2 (F)
Treatments
–0·30·00·30·60·91·21·5
(G)
MCTR
WY
16:0
18:0
c9–18
:1
t10–1
8:1
t11–1
8:118
:2
c9t11
–18:2
t10c1
2–18
:218
:320
:020
:522
:6 MCTR
WY
16:0
18:0
c9–18
:1
t10–1
8:1
t11–1
8:118
:2
c9t11
–18:2
t10c1
2–18
:218
:320
:020
:522
:6–2–101234
c,d,e
d,ec,d,e c,d,e
b,ca
c,d,e
aa,ba,b
b,cb,c
a,b
c
a
b,c
cb
a
b,c,d a,b,c
e
b,c
f
cc
PPAR signalling and glucose synthesis
Immune–related
a
a
Fig. 2. Effect of 150mM of Wy-14 643 (WY) or several long-chain fatty acid (LCFA) treatments for 6 h on the expression of selected genes related to PPARa sig-
nalling and immune response. Mean values with unlike letters were significantly different (P,0·05; Tukey’s corrected). (A) PPARA, PPAR alpha; (B) PPARGC1A,
PPAR gamma coactivator 1-a; (C) ANGPTL4, angiopoietin-like 4; (D) PC, pyruvate carboxylase; (E) IL6; (F) HP, haptoglobin; (G) SAA3, serum amyloid A 3;
(H) SPP1, secreted phosphoprotein 1. M, media; CTR, control (ethanol).
PPAR transcriptomics in bovine cells 183
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
LCFA oxidation
LCFA transport
Infla
mm
atory
resp
onse
TAG synthesis
Cholesterolsynthesis
Signalling
HP
SPP1
ANGPTL4
IL6
SAA3
SOD1
CD36
FABP4FABP3
DBI ACSL1 ACOX1CPT1AACSL5UCP2 SCD LPIN1
DGAT1
HMGCR
HMGCS1
Gluconeogenesis
E, PD
EE
E
EA, F, RB
A, F, PD
A
E
E, PP, RB
RB
RB
RBA, E, PPE
E
I, PD, PP, T, PP
LO
EE
EEE, PD, T
E (2)E, PD, T
E, PP
A, PD, T
E, PD
A, E, T
A, PD, T
PD
E
E
E
E
T
A, E, PDT E, PD, T EE
E E
E, PD
E
E, LO, RBE
PDE E
EE
PC
ACSL4
ACSL3
ACSL6
LPIN2
LPIN3
SREBF1
SREBF2PPARA
PPARGC1A
Wy-14643
Fig. 3. Networks generated with Ingenuity Pathway Analysisw (IPA) among all genes measured with information from the IPA Knowledge Database (IKB; i.e. all
known interactions) (last updated November 2009). The genes are denoted by objects and the letters along the arrows denote the type of effect (activation (A),
effects on gene expression (E), protein–protein interactions (PP), protein–DNA interactions (PD), inhibition (I), RNA binding (RB), effect on translation (T) and
effect on localisation (LO)). Black shade objects fill denotes up-regulation of the gene with Wy-14 643 treatment relative to control. Genes have been grouped
based on main biological function. The bold arrows highlight the downstream targets genes of PPARa uncovered by using the IKB (ACOX1, ACSL1, ANGPL4,
CD36, CPT1A, DBI, FABP3, FABP4, HMGCS1, PPARA, SCD and UCP2) and additional published targets such as ACSL3 and SREBF1 (24), HMGCR (28) and
LPIN1 (30), which overlap with results of our analysis after treatment with Wy-14 643 for 6 h (i.e. similar between non-ruminants and bovine). In dotted lines are
highlighted the PPARa target genes uncovered by using the IKB which were not confirmed by our data. In dashed lines are highlighted positive effect of Wy-
14 643 on expression of SPP1 and LPIN3. Those genes were not recognised in IKB or were not previously published to be PPARa target genes (i.e. novel and
bovine-specific PPARa target genes). The link between PPARA and HMGCS1 uncovered by IPA is actually referring to the link between PPARA and HMGCS2;
however, in rat hepatocytes, HMGCS1 appears to be a PPARa target gene(27). LCFA, long-chain fatty acid; ANGPTL4, angiopoietin-like 4; SPP1, secreted
phosphoprotein 1; HP, haptoglobin; SAA3, serum amyloid A 3; SOD1, superoxide dismutase 1; CD36, CD36 molecule; FABP3 and 4, fatty acid-binding protein 3
and 4; DBI, diazepam binding inhibitor; ACSL1, acyl-coenzyme A synthetase long-chain family member 1; ACOX1, acyl-coenzyme A oxidase 1; CPT1A, carnitine
palmitoyltransferase 1A; ACSL3, 4, 5, and 6, acyl-coenzyme A synthetase long-chain family member 3, 4, 5 and 6; UCP2, uncoupling protein 2; SCD, stearoyl
CoA desaturase; LPIN1, 2 and 3, lipin 1, 2 and 3; DGAT1, diacylglycerol-O-acyltransferase homolog 1; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A
reductase; HMGCS1, 3-hydroxy-3-methylglutaryl-coenzyme A synthase 1; PC, pyruvate carboxylase; SREBF1 and 2, sterol regulatory element binding
transcription factor 1 and 2; PPARA, PPAR alpha; PPARGC1A, PPAR gamma coactivator 1-a.
Table 1. Effect of 150mM treatment with Wy-14 643 (WY) or long-chain fatty acids for 6 h on the expression of genes not reported in Figs. 1 and 2
ACOX1 ACSL3 ACSL4 ACSL5 ACSL6 DBI LPIN2 LPIN3 SOD1 UCP2
Overall P 0·003 , .0001 0·0004 ,0·0001 0·24 0·01 ,0·0001 ,0·0001 ,0·0001 0·002Medium 20·08b,c 20·27d,e,f 20·36c 20·11a,b,c 20·04 0·03 20·21d 20·20d 20·05a,b,c,d 0·02a,b
CTR 20·00a,b,c 20·00e,d 20·00a,b,c 0·00a,b 20·00 20·00 0·00c,d 0·00c,d 0·00a,b,c,d 20·00a,b
WY 0·07a,b,c 0·74b,c 0·07a,b,c 20·05a,b,c 20·01 0·08 0·37b,c 0·59a,b 0·13a 0·13a,b
16 : 0 0·47a 1·95a 0·69a 20·10a,b,c 0·28 0·05 0·92a 0·58a,b 20·05a,b,c,d 20·10a,b
18 : 0 0·28a,b 1·87a 0·63a,b 20·08a,b,c 0·42 0·07 0·81a,b 0·48a,b 20·03a,b,c,d 0·28a
cis9-18 : 1 20·02a,b,c 20·25d,e,f 20·06a,b,c 20·15a,b,c,d 0·00 0·24 0·01c,d 0·29a,b,c 20·02a,b,c,d 0·25a
trans10-18 : 1 0·01a,b,c 0·02c,d,e 0·00a,b,c 20·01a,b 0·04 0·10 0·17c,d 0·22b,c,d 20·18b,c,d 20·05a,b
trans11-18 : 1 20·09b,c 0·03c,d,e 20·29b,c 20·02a,b 0·03 0·41 0·07c,d 0·26a,b,c 20·18b,c,d 20·37b
18 : 2 0·11a,b,c 20·49e,f 0·17a,b,c 20·09a,b,c,d 0·20 0·44 0·14c,d 0·33a,b,c 0·05a,b,c 0·24a
cis9trans11-18 : 2 0·28a,b 0·36b,c,d 0·71a 0·10a 0·44 0·46 0·30c 0·69a 0·11a,b 0·24a
trans10cis12-18 : 2 0·05a,b,c 20·50e,f 0·08a,b,c 20·19a,b,c,d 20·01 0·16 0·06c,d 0·37a,b,c 0·04a,b,c,d 0·21a
18 : 3 20·21c 20·84f 20·55c 20·43d,e 20·17 0·27 0·17c,d 0·23b,c 20·23d 20·13a,b
20 : 0 0·03a,b,c 0·11c,d,e 20·01a,b,c 20·33c,d,e 20·19 0·58 0·03c,d 0·02d,c 20·06a,b,c,d 0·27a
20 : 5 0·02a,b,c 0·96b 0·08a,b,c 20·52e 20·37 0·44 20·16d 20·97d 0·18a 0·14a,b
22 : 6 0·19a,b,c 20·06d,e 0·25a,b,c 20·24b,c,d,e 0·05 0·36 0·20c,d 0·36a,b,c 0·15a 0·28a
SEM 0·09 0·14 0·18 0·06 0·18 0·12 0·09 0·08 0·05 0·10
a,b,c,d,e,f Mean values within a row with unlike superscript letters were significantly different (P,0·05, Tukey’s corrected).ACOX1, acyl-coenzyme A oxidase 1; ACSL3, 4, 5, and 6, acyl-coenzyme A synthetase long-chain family member 3, 4, 5 and 6; DBI, diazepam binding inhibitor (GABA
receptor modulator, acyl-coenzyme A binding protein); LPIN2 and 3, lipin 2 and 3; SOD1, superoxide dismutase 1, soluble; UCP2, uncoupling protein 2 (mitochondrial,proton carrier); CTR, control.
M. Bionaz et al.184
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
Overall, our data confirm that MDBK, and by inference
other bovine cells, respond to PPAR-a agonists and suggest,
based on the thirty genes measured, a peculiar landscape in
the transcriptomics response to WY in this species compared
with non-ruminants. Even though the gene expression land-
scape was different between bovines and non-ruminants,
the activation of PPAR-a appears to strongly control lipid
metabolism as in non-ruminants.
Long-chain fatty acids effect on PPARa activation
It has been well established in non-ruminants that PPAR are
able to bind, and thus are activated by, LCFA. In addition,
LCFA are capable of increasing the expression of PPARA, as
has been shown by treatment of human hepatocytes with
150mM of 16 : 0(40). In non-ruminant species, unsaturated
LCFA are more potent agonists of PPAR isotypes than saturated
LCFA(41–43). The larger degree of PPAR activation in rodents
by polyunsaturated compared with saturated LCFA has been
well established(1).
Our gene expression data (Figs. 1 and 2) summarised by the
gene networks for each single LCFA treatment (Figs. 3– 5 and
S10–S18 in additional file 1, supplementary material for this
article can be found at http://www.journals.cambridge.org/
bjn) suggested that all LCFA elicited their effects partly through
the activation of PPARa (i.e. up-regulated genes positively
affected by WY). The same analyses support a greater
degree of PPAR-a activation by saturated compared with unsa-
turated LCFA.
In the hierarchical clustering analysis considering data with
statistical differences relative to CTR (Fig. 6), responses due to
WY treatment clustered with all LCFA, particularlywith saturated
16 : 0 and 18 : 0 (Fig. 6). Among all treatments EPA appeared to
be the most different. The dendrogram (Fig. 6) highlighted
both a remarkable similarity in effect among saturated LCFA
on the measured genes and a clear separation with the unsatu-
rated LCFA. In this regard, 20 : 0 was the LCFA with the most
similar effect on the expression of the measured genes compa-
red with unsaturated LCFA. Among the saturated LCFA, 16 : 0
and 18 : 0 clustered tightly together. The monounsaturated
Infla
mm
atory
resp
onse
LCFA transport
LCFA oxidation TAG synthesis
Cholesterolsynthesis
Gluconeogenesis
ACSL4
ACSL3
ACSL6
LPIN2
LPIN3SREBF1
SREBF2
Putative TF
PPARA
PPARGC1A
Palmitic acid/stearic acid
Putative TF
SignallingANGPTL4
SPP1
HP
IL6
SAA3
SOD1CD36
CD36
FABP4
FABP3ACOX1CPT1AACSL5
UCP2 SCD
EE E
E, PD
E, PD, T
E, PD
EE
EE
E
E
A, E, RBLO
RB
AA
A
A, E, PP
E
E
RB
I, PD, PP, T, ER
E, PP, RB
EE
E
E E
E
RB
E, PD
E, PP
A, PD, TA, E, T
A, PD, T
PDE, LO, RBE
EE
E
E
E
T
A, E, PDT E, PD, T E
E
LPIN1DGAT1
HMGCR
HMGCS1
PC
DBI ACSL1
Fig. 4. Gene networks encompassing all genes measured that were affected by the saturated long-chain fatty acids (LCFA) 16 : 0 and 18 : 0. The expected biologi-
cal outcome of those two LCFA considering the effect in expression of networks of genes measured would be an increase of lipid metabolism. This would occur
by increasing uptake (CD36) and activation (ACSL1 and ACSL3) of LCFA, by increasing catabolism, particularly in the mitochondria through increase in LCFA
transport (CPT1A), and anabolism, particularly synthesis of TAG (LPIN isoforms and SCD) and cholesterol (HMGCR and HMGCS1). Interestingly, the two satu-
rated LCFA-activated expression of two transcription factors (TF) (PPARGC1A and SREBF1) involved in controlling expression of lipogenic genes and strongly
activated expression of inflammatory response-related genes (IL6, SAA3, and HP) and signalling genes (SPP1 and ANGPTL4). The genes are denoted by objects
and the letters along the arrows denote the type of effect (activation (A), effects on gene expression (E), protein–protein interactions (PP), protein–DNA inter-
actions (PD), inhibition (I), RNA binding (RB), effect on translation (T) and effect on localisation (LO)). Black objects fill denote up-regulation of the gene relative to
control. Genes have been grouped based on main functions. The bolded black lines highlight the downstream targets genes of PPARa (see Fig. 3). In dotted lines
are highlighted genes responsive to Wy-14 643 but not to 16 : 0 and 18 : 0. In dashed lines are highlighted non-ruminants PPARa target genes which expression
was up-regulated by 16 : 0 and 18 : 0 but not by Wy-14 643 treatment. ANGPTL4, angiopoietin-like 4; SPP1, secreted phosphoprotein 1; HP, haptoglobin; SAA3,
serum amyloid A 3; SOD1, superoxide dismutase 1; CD36, CD36 molecule; FABP3 and 4, fatty acid-binding protein 3 and 4; DBI, diazepam binding inhibitor;
ACSL1, acyl-coenzyme A synthetase long-chain family member 1; ACOX1, acyl-coenzyme A oxidase 1; CPT1A, carnitine palmitoyltransferase 1A; ACSL3, 4, 5,
and 6, acyl-coenzyme A synthetase long-chain family member 3, 4, 5 and 6; UCP2, uncoupling protein 2; SCD, stearoyl CoA desaturase; LPIN1, 2 and 3, lipin 1,
2 and 3; DGAT1, diacylglycerol-O-acyltransferase homolog 1; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reductase; HMGCS1, 3-hydroxy-3-methylglutaryl-
coenzyme A synthase 1; PC, pyruvate carboxylase; SREBF1 and 2, sterol regulatory element binding transcription factor 1 and 2; PPARA, PPAR alpha;
PPARGC1A, PPAR gamma coactivator 1-a.
PPAR transcriptomics in bovine cells 185
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
LCFA tended to cluster together as well as the non-
conjugated and conjugated 18 : 2. The latter appeared to have
an effect on the expression of the measured genes, which was
more similar to saturated LCFA and WY than to the monounsa-
turated LCFA. Exogenous c9, t11-CLA clustered together with
20 : 0 and close to WY. Interestingly, if the cluster analysis was
performed without considering the statistical difference in
gene expression relative to CTR, exogenous WY did not cluster
with the saturated LCFA (Fig. S19 in additional file 1, supplemen-
tary material for this article can be found at http://www.journals.
cambridge.org/bjn) but, rather, with unsaturated LCFA and
specifically with DHA. The correlation analysis (additional file
3, supplementary material for this article can be found at
http://www.journals.cambridge.org/bjn) using both significant
and non-significant outcomes seems to support a close relation-
ship between WY and DHA. Even though the transcriptomics
effect of all treatments correlated significantly, the lowest
correlation was observed for 16 : 0 and 18 : 0 with other
LCFA and WY. The clustering analysis among genes (Fig. 6)
demonstrated a large similarity between the genes involved in
LCFA entry into cells (CD36) and into the mitochondria
(CPT1A). Interestingly, the two signalling molecules analysed
(ANGPTL4 and SPP1) also tended to cluster together. No
other functional clusters were observed among the measured
genes.
The hierarchical clustering underscored that the response to
palmitic and stearic acids in MDBK was very similar to WY,
suggesting those LCFA being the more potent PPAR-a agonists
among the one tested; however, the gene expression data
(with the sole exception of FABP4 and ANGPTL4) showed a
stronger response for nearly all the PPAR-a target genes
measured with the two saturated LCFA compared with WY.
It is noteworthy that the expression of HMGCS1, which is a
known PPAR-a target in non-ruminants, was only numerically
up-regulated by WY (i.e. it is a weak PPAR-a target in rumi-
nants), but was significantly up-regulated in response to the
two saturated LCFA (Fig. 2). The apparently greater potency
of saturated LCFA, particularly 16 : 0, to activate PPAR-a in
MDBK has also been observed previously(10,11). Recently,
we observed a larger response with saturated, namely 16 : 0,
Infla
mm
atory
resp
onse
LCFA transport
LCFA oxidation TAG synthesis
Cholesterolsynthesis
Gluconeogenesis
ACSL4
ACSL3
ACSL6
LPIN2
LPIN3
Putative TF
SREBF2PPARA
Putative TF
SPP1IL6
HP
SAA3
SOD1
CD36
FABP4
DBIFABP3
ACOX1CPT1AACSL5UCP2 SCD LPIN1
DGAT1
HMGCR
HMGCS1
PC
EE
E
E
E
EE
A, I
RBLO
E
E
A, I
E
E
E
E
E
T
EPD
EE
E, LO, RB
A, E, PDE, PD, T E E
E E
E, PD
E, PD, TA, PD, TE, PP
E, PD, TE, PDE
A, E, T
A, PD, T
RB
A, E, PPI, PD, PP, T, PP
A, E, RBE, PP, RB
RB
E PD
EBPD
T
EE
EE
EE
EE
ACSL1
ANGPTL4Signalling
PPARGC1A
5, 8, 11, 14, 17-EPA
SREBF1
Fig. 5. Gene networks encompassing all genes measured that were affected by EPA among all genes measured. The network analysis among measured genes
indicated that treatment with EPA probably induced long-chain fatty acid (LCFA) oxidation in mitochondria, TAG and cholesterol synthesis, and production of sig-
nalling molecules; however, some of the genes involved in lipid metabolism were down-regulated (e.g. DGAT1 and ACSL5) as well PC and IL6, with a likely
decrease in activation of IL6 network genes. In this regard, it was noteworthy the observed up-regulation of expression of HP, which probably indicates that
expression of this gene was induced by EPA through a network not involving IL6. The genes are denoted by objects and the letters along the arrows denote the
type of effect (activation (A), effects on gene expression (E), protein–protein interactions (PP), protein–DNA interactions (PD), inhibition (I), RNA binding (RB),
effect on translation (T) and effect on localisation (LO)). Black objects fill denotes up-regulation and grey down-regulation of the gene relative to control. Genes
have been grouped based on main functions. The bold lines highlight the downstream targets genes of PPARa that overlap with genes up-regulated by EPA treat-
ment (see Fig. 3). In dashed lines are highlighted genes affected by EPA probably through other (putative) transcription factors (TF), among those all except HP
were down-regulated and HMGCS1 was not up-regulated by Wy-14 643 but was up-regulated by EPA treatment. The dotted lines highlighted the lack of effect of
EPA on ACSL1 and LPIN3, the only ruminant-specific PPARa target genes (see Fig. 3) no affected by EPA. ANGPTL4, angiopoietin-like 4; SPP1, secreted phos-
phoprotein 1; HP, haptoglobin; SAA3, serum amyloid A 3; SOD1, superoxide dismutase 1; CD36, CD36 molecule; FABP3 and 4, fatty acid-binding protein 3 and
4; DBI, diazepam binding inhibitor; ACSL1, acyl-coenzyme A synthetase long-chain family member 1; ACOX1, acyl-coenzyme A oxidase 1; CPT1A, carnitine pal-
mitoyltransferase 1A; ACSL3, 4, 5, and 6, acyl-coenzyme A synthetase long-chain family member 3, 4, 5 and 6; UCP2, uncoupling protein 2; SCD, stearoyl CoA
desaturase; LPIN1, 2 and 3, lipin 1, 2 and 3; DGAT1, diacylglycerol-O-acyltransferase homolog 1; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reductase;
HMGCS1, 3-hydroxy-3-methylglutaryl-coenzyme A synthase 1; PC, pyruvate carboxylase; SREBF1 and 2, sterol regulatory element binding transcription factor 1
and 2; PPARA, PPAR alpha; PPARGC1A, PPAR gamma coactivator 1-a.
M. Bionaz et al.186
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
compared with a specific PPAR-g agonist in increasing the
expression of several putative PPAR-g target genes in MacT
cells(34). From a physiological point of view, those data
appear to highlight a teleological evolutionary adaptation of
these nuclear receptors in ruminants to greater availability of
saturated rather than unsaturated LCFA. The major LCFA in
ruminant blood under most circumstances are palmitic, stearic
and oleic acid(44). In addition, the lower response of PPAR-a in
MDBK to polyunsaturated LCFA, with the exception of EPA, is
meaningful considering the large degree of ruminal biohydro-
genation of those LCFA in bovine(45). It is interesting from this
point of view that, among unsaturated LCFA, the more potent
activator of PPAR-a after EPA was c9, t11-CLA (Fig. S11 in
additional file 1, supplementary material for this article can
be found at http://www.journals.cambridge.org/bjn), which
can accumulate in the rumen due to unsaturated lipid feed-
ing(46) or can be synthesised endogenously from vaccenic
acid (another biohydrogenation intermediate) in tissues via
the enzyme SCD(47). This LCFA is one of the most abundant
conjugated 18 : 2 in the blood of dairy cows(46) and the most
abundant CLA in ruminant milk and meat(47).
The overall PPRE analysis (Table S7 in additional file 1, sup-
plementary material for this article can be found at http://
www.journals.cambridge.org/bjn), but especially the corre-
lation analysis between gene expression and the in silico
PPRE binding strength (additional file 3, supplementary
material for this article can be found at http://www.journals.
cambridge.org/bjn), does not provide evidence of greater
potential for LCFA in bovine to activate PPAR-g and PPAR-b/
d compared with PPAR-a. In non-ruminants, all three PPAR
subtypes appear able to bind and be activated by LCFA(48).
Differences exist in non-ruminants in the binding capacity of
the three PPAR isotypes, with a and b/d being equally able
to bind saturated and unsaturated LCFA while PPAR-g appears
to be more prone to binding polyunsaturated LCFA(48). The
PPRE analysis of the three bovine PPAR isotypes seems to indi-
cate that the LCFA could have increased/decreased expression
of measured genes also through PPAR-g. This was surmised
because its PPRE is the most abundant in the genomic
sequences of the genes analysed (Table S7 and Fig. S4 in
additional file 1, supplementary material for this article can
be found at http://www.journals.cambridge.org/bjn). In this
regard, PPAR-b/d appears to be a weaker player and our
data do not support a strong functional overlap with PPAR-a.
To evaluate the potential differences in LCFA binding in
bovine cells, we have conducted an initial analysis of the
three-dimensional structure of the ligand-binding domain of
the three bovine PPAR isotypes (Fig. S20 in additional file 1,
supplementary material for this article can be found at
http://www.journals.cambridge.org/bjn) which does allow,
with limitations, to make some inferences about the difference
in binding capacity for LCFA between the three PPAR. In this
1:1 1·5–1·5
20:0
5
16:0
0
18:0
0
WY
20:0
0
c9t1
1–18
:2
t10c
12–1
8:2
c9-1
8:1
t10–
18:1
t11–
18:1
18:0
3
22:0
6
18:0
2
CD36ANGPTL4CPT1ASPP1LPIN1FABP4HMGCRSAA1SCDSREBF1ACSL1PPARGC1AACSL3HMGCS1IL6LPIN3HPLPIN2ACSL4FABP3ACSL5PCDGAT1
Fig. 6. Hierarchical clustering of gene expression data between treatments and between genes considering only data with significant differences relative to the
control (CTR) using Genesis software(64). Hierarchical cluster using all genes without considering the significance is reported in Fig. S19 in additional file 1 (sup-
plementary material available online at http://www.journals.cambridge.org/bjn). The dendrogram allows visualisation of clusters of similarity in expression pattern
between treatments (links denoted by the lines at the top of the figure) and between genes (links denoted by the lines at the left side of the picture). Log2 fold-
change in expression relative to CTR are denoted by shades of black, increase; light grey-white, down-regulated; grey, no change relative to CTR according to
the intensity bar at the top of the Fig. (refer to Table 1 and Figs. 1 and 2 for statistical differences). White dots denote the largest responses (up or down) in
mRNA expression relative to CTR for each gene. CD36, CD36 molecule; ANGPTL4, angiopoietin-like 4; CPT1A, carnitine palmitoyltransferase 1A; SPP1,
secreted phosphoprotein 1; LPIN1, lipin 1; FABP 4, fatty acid-binding protein 4; HMGCR, 3-hydroxy-3-methylglutaryl-coenzyme A reductase; SAA1, serum
amyloid A 1; SCD, stearoyl CoA desaturase; SREBF1, sterol regulatory element binding transcription factor 1; ACSL1, acyl-coenzyme A synthetase long-chain
family member 1; PPARGC1A, PPAR gamma coactivator 1-a; ACSL3, acyl-coenzyme A synthetase long-chain family member 3; HMGCS1, 3-hydroxy-3-methyl-
glutaryl-coenzyme A synthase 1; LPIN3, lipin 3; HP, haptoglobin; LPIN2, lipin 2; ACSL4, acyl-coenzyme A synthetase long-chain family member 4; FABP3,
fatty acid-binding protein 3; ACSL5, acyl-coenzyme A synthetase long-chain family member 5; PC, pyruvate carboxylase; DGAT1, diacylglycerol-O-acyltransferase
homolog 1.
PPAR transcriptomics in bovine cells 187
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
regard, it was interesting from the three-dimensional images
the apparent larger and more neutrally charged ligand
pocket in PPAR-a compared with the other two PPAR isotypes,
which is suggestive that it can accommodate larger and more
neutral molecules (larger neutral carbon chain?) compared
with other PPAR. The more negative charge in PPAR-g and
the very large positive charge in PPAR-b/d also emerged as
interesting characteristics of the differences among PPAR iso-
types. As stated above, the in silico data have to be considered
preliminary and in-depth molecular analyses (e.g. crystallo-
graphy) are needed in order to uncover specific differences
in the ligand-binding capacity between the three bovine
PPAR isotypes.
Gene networks affected by long-chain fatty acids
Most of the measured genes, with the exception of ACOX1,
ACSL4, ACSL6, DBI, SOD1 and UCP2, were significantly
affected after treatment with one or more LCFA (Figs. 1 and 2
and Table 1). The saturated LCFA appeared to have had a
larger effect on the expression of the measured genes com-
pared with unsaturated LCFA. The largest effect was observed
for 16 : 0 and 18 : 0 (up-regulation of seventeen out of thirty
measured transcripts or 56·7 %, Fig. 4) among the saturated
(20 : 0 affected 30 % of genes, with four down-regulated, Fig.
S10 in additional file 1, supplementary material for this article
can be found at http://www.journals.cambridge.org/bjn) and
EPA among the unsaturated LCFA (60 % measured transcripts
affected, six genes down-regulated, Fig. 5). Among the
other unsaturated LCFA, the effect on the expression of
measured genes was c9, t11CLA $ 18 : 2 . t10, c12CLA ¼
DHA . .18 : 3 $ c9-18 : 1 ¼ t11- $ t10-18 : 1 (Figs. S10–S18
in additional file 1, supplementary material for this article
can be found at http://www.journals.cambridge.org/bjn).
When considering the gene networks among the measured
genes, all the LCFA tested appeared capable of eliciting
oxidation of fatty acids ( * CD36 and * CPT1A) and signalling
through ANGPTL4 (summary in Figs. 4 and 5 and S10–S18 in
additional file 1, supplementary material for this article can
be found at http://www.journals.cambridge.org/bjn). This
general response was probably due to the fact that the
culture media used was devoid of LCFA. Before starting the
experiment, the MDBK cells were in the media with 10 %
fetal bovine serum containing approximately 0·015 mEq/l
NEFA(11). This was approximately 10-fold less than the
plasma NEFA concentration in bovines under normal con-
ditions and approximately 50-fold less compared with cows
just after parturition (see, for instance, Bionaz et al.(49)). Inter-
estingly, CD36 and ANGPTL4 were the least abundant genes
among all those measured in MDBK (Fig. S1 in additional
file 1, supplementary material for this article can be found at
http://www.journals.cambridge.org/bjn).
The effect on ANGPTL4 is of interest because this gene
codes for a protein that is synthesised in several tissues includ-
ing adipose and liver and seems to play roles within the liver
and peripheral tissues including adipose(50). For example,
ANGPTL4 is up-regulated during fasting in bovines(51) and
non-ruminants(50), and can bind and inactivate lipoprotein
lipase in adipose tissue. An end-result of ANGPTL4 action is
to increase plasma TAG and cholesteryl ester and decrease
the uptake of fatty acids and cholesterol into peripheral tissues
but, probably, increase availability of NEFA to the liver(50).
Interestingly, it was demonstrated recently that ANGPTL4 is
up-regulated during the lipopolysacharide challenge in
murine adipose, muscle and heart(52). Thus, ANGPTL4 can
be considered a novel positive acute-phase protein(52). In
the same study, the transcription of ANGPTL4 in the liver
was down-regulated during the early response and up-
regulated during the late response to lipopolysacharide. In
the present study, ANGPTL4 responded rapidly regardless of
the type of LCFA treatment. We cannot exclude that in some
of the treatments (e.g. saturated LCFA) ANGPTL4 responded
as an acute-phase response molecule.
From an in vivo ruminant perspective, the responses in
gene expression observed in the present study can be classi-
fied into two categories, those associated with the LCFA
found primarily in the diet (16 : 0, 18 : 0, c9-18 : 1, 18 : 2,
18 : 3, EPA and DHA) and those associated with ruminal
metabolism (t10-18 : 1, t11-18 : 1, c9, t11-CLA, t10,c12-CLA
and 20 : 0). An in-depth discussion of the gene networks
affected by those categories of LCFA is reported in additional
file 1 (supplementary material available online at http://www.
journals.cambridge.org/bjn). In the following section, we
briefly report the main findings within each category.
Dietary long-chain fatty acids
Among dietary LCFA, 16 : 0 and 18 : 0 and EPA elicited the
largest changes in gene expression (Figs. 1 and 2). All three
LCFA appear to have increased both catabolic and anabolic
utilisation of LCFA (Figs. 4 and 5). Results for the saturated
LCFA confirmed previous data obtained in hepatocytes from
pre-ruminant calves(53). Overall, the data indicated that satu-
rated LCFA and EPA would increase uptake and utilisation of
NEFA along with increased formation of cholesterol. The
increase in cholesterol synthesis can be considered a positive
outcome in bovine liver (and particularly in the peripartal
period) because it is essential for the formation of lipoproteins
to remove TAG(54). The present and previous(53) data appear
to be supported by a recent in vivo study in which greater
plasma cholesterol level was measured in late pregnant
cows fed a palmitic acid-enriched diet(55).
The difference in the regulation of the expression of pro-
inflammatory genes between the two saturated fatty acids
and EPA was noteworthy (Fig. 2). The saturated LCFA elicited
an evident increase in inflammation, at least considering the
expression of the measured genes; EPA appeared to have
reduced (mostly through + IL6) inflammation. Although this
represents the first evidence in bovine cells, judging from
non-ruminant data, the decrease of inflammation by EPA(56)
and increase of inflammation by 16 : 0 are not entirely novel
findings(57–59). The inflammatory response induced by palmi-
tate appears to occur through the activation of the NF-kB
transcription factor, which, in turn, increases the expression
of IL6 (57–59). In our case, the ‘putative transcription factor(s)’
M. Bionaz et al.188
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
reported in Fig. 4 could probably be NF-kB; however, this
remains to be proven experimentally in bovines.
Even though PPARA was not up-regulated significantly by
the dietary LCFA in the present study, up-regulation of its
known co-activators, PPARGC1A (26) and LPIN1 (19), was sub-
stantial in response to 16 : 0, 18 : 0 and EPA (Figs. 1 and 2),
suggesting that, besides activating PPAR-a through direct bind-
ing, those LCFA also increase its activation by enhancing the
availability of co-activators. The other dietary LCFA (c9-18 : 1,
18 : 2, 18 : 3 and DHA) elicited only moderate effects on
measured genes (Figs. 1 and 2 and S13–S15 in additional
file 1, supplementary material for this article can be found at
http://www.journals.cambridge.org/bjn).
Rumen-related long-chain fatty acids
Far fewer changes were observed with LCFA that arise from
ruminal metabolism (Figs. 1 and 2). Of all rumen-derived
fatty acids investigated, 20 : 0 and c9,t11CLA elicited the
most changes in gene expression (Figs. S10 and S11 in
additional file 1, supplementary material for this article can
be found at http://www.journals.cambridge.org/bjn). Phytanic
acid (20 : 0) is a branched-chained LCFA derived from the rum-
inal metabolism of chlorophyll(60) and, as such, its production
and deposition in tissues or milk is probably greater in cattle
fed forage-based diets(61). Phytanic acid is known to be a
potent murine liver PPAR-a ligand(13) but also activates the
other two PPAR isoforms, at least in the rat(62). The effect of
20 : 0 on gene expression of measured genes suggested a
reduced TAG and cholesterol synthesis in MDBK (Figs. S10
in additional file 1, supplementary material for this article
can be found at http://www.journals.cambridge.org/bjn).
To our knowledge, no previous bovine cell culture work
has been carried out with 20 : 0, but in rat hepatocytes, it
was demonstrated that it is involved in increasing glucose
metabolism(62).
The functional responses of measured genes with exo-
genous c9,t11CLA (or rumenic acid) as inferred by network
analysis (Fig. S11, supplementary material for this article can
be found at http://www.journals.cambridge.org/bjn) is partly
consistent with previous findings. In fact, as all treatments in
the present study, c9,t11CLA increased the expression of
genes associated with the oxidation of LCFA, which is com-
monly observed in rodents and humans fed with CLA(63).
The other rumen-related unsaturated LCFA had very modest
effects on the expression of measured genes (Figs. 1 and 2
and summary in Figs. S12 and S16–S18 with discussion in
additional file 1, supplementary material for this article can
be found at http://www.journals.cambridge.org/bjn).
In a recent study, we investigated the transcriptomics effect
of several LCFA on MacT cells(34). Several of those LCFA were
also tested in the present study. To evaluate consistency in
bovine cell responses to LCFA, we have conducted a compari-
son between the gene expression results between MDBK and
MacT cells. The results are summarised in Table S10
(additional file 1 with therein relative discussion, available
online at http://www.journals.cambridge.org/bjn). The anal-
ysis uncovered differences in response to LCFA between the
two immortalised bovine cells, with a strikingly consistent
greater response associated with saturated rather than
unsaturated LCFA. The results from this comparison suggest
caution when inferring data from one cell type to another;
however, results support the fact that, opposite to non-rumi-
nants, ruminant cells are more sensitive to saturated rather
than unsaturated LCFA. The large response of bovine cells to
PPAR agonists and LCFA strongly supports an effect of LCFA
on the transcriptome.
Conclusions
The present study uncovered that ACSL1, ACSL3, ANGPTL4,
CD36, CPT1A, FABP4, HMGCR, LPIN1, SCD and SREBF1,
which are known PPAR-a targets in non-ruminants, are also
induced by WY in the bovine. Novel and apparently bovine-
specific PPAR-a targets were SPP1and LPIN3.
Our data provided potential avenues for the future, and
more oriented, experiments in order to test the feasibility of
using dietary LCFA to finely modulate metabolism in ruminant
tissues such as the liver. The present study provides support
for several LCFA as being PPAR-a agonists (particularly 16 : 0,
18 : 0 and EPA), but additional studies should be conducted
to examine with greater confidence the binding and potency
of activation of PPAR-a by LCFA, e.g. studies using the same
experimental design but with the use of PPAR-a antagonists
or the inhibition of PPAR-a expression/translation (e.g. by
using siRNA).
In conclusion, the results from the present study strongly
support the possibility that dietary LCFA, and particularly
16 : 0 and 18 : 0, are able to modulate ruminant metabolism,
particularly lipid metabolism, with the major effects probably
induced via the activation of PPAR. Those findings need
to be verified in an in vivo milieu with the exciting possi-
bility that, if verified, they will open novel opportunities
for fine-tuning the regulation of bovine metabolism via care-
ful/controlled dietary approaches. Those could have a tre-
mendous impact in the dairy industry, for example, by
providing the means to prevent metabolic disorders such
as fatty liver.
Acknowledgements
We gratefully acknowledge funding from National Institute
of Food and Agriculture (Washington, DC, USA) under
projects ILLU-538-952 and ILLU-538-961. M. B. and J. J. L.
conceived and designed the study. As part of the MS
thesis B. J. T. coordinated and performed the study and
qPCR analysis, and conducted the initial statistical analysis
of qPCR data. M. B. helped perform and coordinate the
study, performed final statistical analysis of the qPCR
data, performed the PPRE and three-dimensional analyses,
prepared the figures, and wrote the manuscript. J. J. L.
supervised all aspects of the study and helped write the
final manuscript. All authors read and approved the final
manuscript. There are no financial or other contractual
agreements that might cause conflict of interests.
PPAR transcriptomics in bovine cells 189
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
References
1. Bragt MC & Popeijus HE (2008) Peroxisome proliferator-activated receptors and the metabolic syndrome. PhysiolBehav 94, 187–197.
2. Kumamoto T & Ide T (1998) Comparative effects of alpha-and gamma-linolenic acids on rat liver fatty acid oxidation.Lipids 33, 647–654.
3. Desvergne B, Michalik L & Wahli W (2006) Transcriptionalregulation of metabolism. Physiol Rev 86, 465–514.
4. Jay MA & Ren J (2007) Peroxisome proliferator-activatedreceptor (PPAR) in metabolic syndrome and type 2 diabetesmellitus. Curr Diabetes Rev 3, 33–39.
5. Kim MS, Sweeney TR, Shigenaga JK, et al. (2007) Tumornecrosis factor and interleukin 1 decrease RXRalpha, PPAR-alpha, PPARgamma, LXRalpha, and the coactivators SRC-1,PGC-1alpha, and PGC-1beta in liver cells. Metabolism 56,267–279.
6. Zandbergen F & Plutzky J (2007) PPARalpha in atherosclero-sis and inflammation. Biochim Biophys Acta 1771, 972–982.
7. Drackley JK (1999) ADSA Foundation Scholar Award.Biology of dairy cows during the transition period: thefinal frontier? J Dairy Sci 82, 2259–2273.
8. Loor JJ, Dann HM, Everts RE, et al. (2005) Temporal geneexpression profiling of liver from periparturient dairy cowsreveals complex adaptive mechanisms in hepatic function.Physiol Genomics 23, 217–226.
9. Selberg KT, Staples CR, Luchini ND, et al. (2005) Dietarytrans octadecenoic acids upregulate the liver gene encodingperoxisome proliferator-activated receptor-alpha in tran-sition dairy cows. J Dairy Res 72, 107–114.
10. Bionaz M, Baumrucker CR, Shirk E, et al. (2008) Characteriz-ation of madin-darby bovine kidney cell line for peroxisomeproliferator-activated receptors: temporal response andsensitivity to fatty acids. J Dairy Sci 91, 2808–2813.
11. Thering BJ, Bionaz M & Loor JJ (2009) Long-chain fattyacid effects on peroxisome proliferator-activated receptor-alpha-regulated genes in Madin-Darby bovine kidney cells:optimization of culture conditions using palmitate. J DairySci 92, 2027–2037.
12. Litherland NB, Wallace RL, Bionaz M, et al. (2010) Effectsof peroxisome proliferator activated receptor-a agonistsclofibrate and fish oil on liver fatty acid metabolism inweaned dairy calves. J Dairy Sci 93, 2404–2418.
13. Wolfrum C, Borrmann CM, Borchers T, et al. (2001) Fattyacids and hypolipidemic drugs regulate peroxisome prolif-erator-activated receptors alpha- and gamma-mediatedgene expression via liver fatty acid binding protein: a signal-ing path to the nucleus. Proc Natl Acad Sci U S A 98,2323–2328.
14. GoyaK, Sumitani S, XuX, et al. (2004) Peroxisomeproliferator-activated receptor alpha agonists increase nitric oxidesynthase expression in vascular endothelial cells. ArteriosclerThromb Vasc Biol 24, 658–663.
15. Hayashida K, Kume N, Minami M, et al. (2004) Peroxisomeproliferator-activated receptor alpha ligands activate tran-scription of lectin-like oxidized low density lipoproteinreceptor-1 gene through GC box motif. Biochem BiophysRes Commun 323, 1116–1123.
16. Bionaz M & Loor JJ (2008) Gene networks driving bovinemilk fat synthesis during the lactation cycle. BMC Genomics9, 366.
17. Durgan DJ, Smith JK, Hotze MA, et al. (2006) Distincttranscriptional regulation of long-chain acyl-CoA synthetaseisoforms and cytosolic thioesterase 1 in the rodent heart by
fatty acids and insulin. Am J Physiol Heart Circ Physiol290, H2480–H2497.
18. Phan J, Peterfy M & Reue K (2004) Lipin expression preced-ing peroxisome proliferator-activated receptor-gamma iscritical for adipogenesis in vivo and in vitro. J Biol Chem279, 29558–29564.
19. Finck BN, Gropler MC, Chen Z, et al. (2006) Lipin 1 is aninducible amplifier of the hepatic PGC-1alpha/PPARalpharegulatory pathway. Cell Metab 4, 199–210.
20. Lefever S, Hellemans J, Pattyn F, et al. (2009) RDML:structured language and reporting guidelines for real-timequantitative PCR data. Nucleic Acids Res 37, 2065–2069.
21. Bustin SA, Benes V, Garson JA, et al. (2009) The MIQE guide-lines: minimum information for publication of quantitativereal-time PCR experiments. Clin Chem 55, 611–622.
22. Heinaniemi M, Uski JO, Degenhardt T, et al. (2007)Meta-analysis of primary target genes of peroxisome prolif-erator-activated receptors. Genome Biol 8, R147.
23. Guex N & Peitsch MC (1997) SWISS-MODEL and theSwiss-PdbViewer: an environment for comparative proteinmodeling. Electrophoresis 18, 2714–2723.
24. Rakhshandehroo M, Hooiveld G, Muller M, et al. (2009)Comparative analysis of gene regulation by the transcriptionfactor PPARa between mouse and human. PLoS One 4,e6796.
25. Lehmann JM, Lenhard JM, Oliver BB, et al. (1997)Peroxisome proliferator-activated receptors alpha and gammaare activated by indomethacin and other non-steroidalanti-inflammatory drugs. J Biol Chem 272, 3406–3410.
26. Mandard S, Muller M & Kersten S (2004) Peroxisome prolif-erator-activated receptor alpha target genes. Cell Mol Life Sci61, 393–416.
27. Tamura K, Ono A, Miyagishima T, et al. (2006) Profilingof gene expression in rat liver and rat primary culturedhepatocytes treated with peroxisome proliferators. J ToxicolSci 31, 471–490.
28. van der Meer DL, Degenhardt T, Vaisanen S, et al. (2010)Profiling of promoter occupancy by PPAR{alpha} in humanhepatoma cells via ChIP-chip analysis. Nucleic Acids Res38, 2839–2850.
29. Pineda Torra I, Jamshidi Y, Flavell DM, et al. (2002) Charac-terization of the human PPARalpha promoter: identificationof a functional nuclear receptor response element. MolEndocrinol 16, 1013–1028.
30. Reue K & Zhang P (2008) The lipin protein family: dual rolesin lipid biosynthesis and gene expression. FEBS Lett 582,90–96.
31. Wang Y, Botolin D, Xu J, et al. (2006) Regulation of hepaticfatty acid elongase and desaturase expression in diabetesand obesity. J Lipid Res 47, 2028–2041.
32. Morandini P, Salamini F & Gantet P (2005) Engineeringof plant metabolism for drug and food. Curr Med ChemImmunol Endocr Metab Agents 5, 103–112.
33. Horton JD (2002) Sterol regulatory element-binding proteins:transcriptional activators of lipid synthesis. Biochem SocTrans 30, 1091–1095.
34. Kadegowda AK, Bionaz M, Piperova LS, et al. (2009) Peroxi-some proliferator-activated receptor-gamma activation andlong-chain fatty acids alter lipogenic gene networks inbovine mammary epithelial cells to various extents. J DairySci 92, 4276–4289.
35. Lambe KG, Woodyatt NJ, Macdonald N, et al. (1999) Speciesdifferences in sequence and activity of the peroxisomeproliferator response element (PPRE) within the acyl CoAoxidase gene promoter. Toxicol Lett 110, 119–127.
M. Bionaz et al.190
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777
36. Rodriguez C, Noe V, Cabrero A, et al. (2000) Differences inthe formation of PPARalpha-RXR/acoPPRE complexesbetween responsive and nonresponsive species upon fibrateadministration. Mol Pharmacol 58, 185–193.
37. Lee JH, Banerjee A, Ueno Y, et al. (2008) Potential relation-ship between hepatobiliary osteopontin and peroxisomeproliferator-activated receptor alpha expression followingethanol-associated hepatic injury in vivo and in vitro. ToxicolSci 106, 290–299.
38. Oyama Y, Akuzawa N, Nagai R, et al. (2002) PPARgammaligand inhibits osteopontin gene expression through interfer-ence with binding of nuclear factors to A/T-rich sequence inTHP-1 cells. Circ Res 90, 348–355.
39. Forman BM, Tontonoz P, Chen J, et al. (1995) 15-Deoxy-delta 12, 14-prostaglandin J2 is a ligand for the adipocytedetermination factor PPAR gamma. Cell 83, 803–812.
40. Swagell CD, Henly DC & Morris CP (2007) Regulation ofhuman hepatocyte gene expression by fatty acids. BiochemBiophys Res Commun 362, 374–380.
41. Kliewer SA, Sundseth SS, Jones SA, et al. (1997) Fatty acidsand eicosanoids regulate gene expression through directinteractions with peroxisome proliferator-activated receptorsalpha and gamma. Proc Natl Acad Sci U S A 94, 4318–4323.
42. Desvergne B & Wahli W (1999) Peroxisome proliferator-activated receptors: nuclear control of metabolism. EndocrRev 20, 649–688.
43. Larter CZ (2007) Not all models of fatty liver are createdequal: understanding mechanisms of steatosis developmentis important. J Gastroenterol Hepatol 22, 1353–1354.
44. Leroy JL, Vanholder T, Mateusen B, et al. (2005) Non-esterified fatty acids in follicular fluid of dairy cows andtheir effect on developmental capacity of bovine oocytesin vitro. Reproduction 130, 485–495.
45. Hess BW, Moss GE & Rule DC (2008) A decade of develop-ments in the area of fat supplementation research with beefcattle and sheep. J Anim Sci 86, E188–E204.
46. Loor JJ, Ferlay A, Ollier A, et al. (2005) High-concentratediets and polyunsaturated oils alter trans and conjugatedisomers in bovine rumen, blood, and milk. J Dairy Sci 88,3986–3999.
47. Dhiman TR, Nam SH & Ure AL (2005) Factors affecting con-jugated linoleic acid content in milk and meat. Crit Rev FoodSci Nutr 45, 463–482.
48. Grimaldi PA (2007) Peroxisome proliferator-activated recep-tors as sensors of fatty acids and derivatives. Cell Mol Life Sci64, 2459–2464.
49. Bionaz M, Trevisi E, Calamari L, et al. (2007) Plasma para-oxonase, health, inflammatory conditions, and liver functionin transition dairy cows. J Dairy Sci 90, 1740–1750.
50. Lichtenstein L & Kersten S (2010) Modulation of plasmaTG lipolysis by Angiopoietin-like proteins and GPIHBP1.Biochim Biophys Acta 1801, 415–420.
51. Loor JJ, Everts RE, Bionaz M, et al. (2007) Nutrition-inducedketosis alters metabolic and signaling gene networks in liverof periparturient dairy cows. Physiol Genomics 32, 105–116.
52. Lu B, Moser A, Shigenaga JK, et al. (2010) The acute phaseresponse stimulates the expression of angiopoietin likeprotein 4. Biochem Biophys Res Commun 391, 1737–1741.
53. Mashek DG & Grummer RR (2003) Effects of long chain fattyacids on lipid and glucose metabolism in monolayer culturesof bovine hepatocytes. J Dairy Sci 86, 2390–2396.
54. Hocquette JF & Bauchart D (1999) Intestinal absorption,blood transport and hepatic and muscle metabolism offatty acids in preruminant and ruminant animals. ReprodNutr Dev 39, 27–48.
55. Andersen JB, Ridder C & Larsen T (2008) Priming the cowfor mobilization in the periparturient period: effects ofsupplementing the dry cow with saturated fat or linseed.J Dairy Sci 91, 1029–1043.
56. Soumura M, Kume S, Isshiki K, et al. (2010) Oleate and eico-sapentaenoic acid attenuate palmitate-induced inflammationand apoptosis in renal proximal tubular cell. BiochemBiophys Res Commun 402, 265–271.
57. Ajuwon KM & Spurlock ME (2005) Palmitate activatesthe NF-kappaB transcription factor and induces IL-6 andTNFalpha expression in 3T3-L1 adipocytes. J Nutr 135,1841–1846.
58. Ciapaite J, van Bezu J, van Eikenhorst G, et al. (2007)Palmitate and oleate have distinct effects on the inflamma-tory phenotype of human endothelial cells. Biochim BiophysActa 1771, 147–154.
59. Coll T, Eyre E, Rodriguez-Calvo R, et al. (2008) Oleatereverses palmitate-induced insulin resistance and inflam-mation in skeletal muscle cells. J Biol Chem 283,11107–11116.
60. Dawson RM, Hemington N, Grime D, et al. (1974) Lipolysisand hydrogenation of galactolipids and the accumulationof phytanic acid in the rumen. Biochem J 144, 169–171.
61. Lee MR, Connelly PL, Tweed JK, et al. (2006) Effects of high-sugar ryegrass silage and mixtures with red clover silage onruminant digestion. 2. Lipids. J Anim Sci 84, 3061–3070.
62. Heim M, Johnson J, Boess F, et al. (2002) Phytanic acid, anatural peroxisome proliferator-activated receptor (PPAR)agonist, regulates glucose metabolism in rat primary hepato-cytes. FASEB J 16, 718–720.
63. DeLany JP & West DB (2000) Changes in body compositionwith conjugated linoleic acid. J Am Coll Nutr 19, 487S–493S.
64. Sturn A, Quackenbush J & Trajanoski Z (2002) Genesis: clus-ter analysis of microarray data. Bioinformatics 18, 207–208.
PPAR transcriptomics in bovine cells 191
British
Journal
ofNutrition
Dow
nloaded from https://w
ww
.cambridge.org/core . IP address: 54.39.106.173 , on 12 Jul 2020 at 17:45:07 , subject to the Cam
bridge Core terms of use, available at https://w
ww
.cambridge.org/core/term
s . https://doi.org/10.1017/S0007114511002777